SU-FF-T-659: CRADLE - A Software Architecture for Radiation Therapy Treatment Planning

Abstract

Purpose: The treatment planning process has not changed in 20 years: patient data is imported, “invisible” data is created inside the treatment planning system (TPS) to describe the treatment course, and data is exported to the treatment device. An oncology information system manages this exported data and has become the center of the modern department, while the TPS has remained largely static. We deem it essential to incorporate new capabilities, such internal data management, into the TPS. Method and Materials: We present a novel architecture (implemented as CRADLE) that addresses three problems: (1) application data management congruent with DICOM, (2) user‐interface management based on the immediate mode model (IMGUI), and (3) fine‐grained state and version data management. We demonstrate CRADLE in our proton scanning application. Results: DICOM defines a model for RT data associations, but is only used between application entities. We internalize this model and design our data associations congruent with it and achieve a near‐transparent mapping between internal and external representations. Current user‐interface (UI) designs use an event call‐back model. This model moves authority from the application to the UI‐code (created by builders), requires user synchronization between application and UI‐code data, and is characterized by a static, UI‐code controlled, environment. In contrast, IMGUI moves authority and data into the application, achieves local UI and context control and allows for dynamic UI control. Finally, TPS data management must achieve fine‐grained version control to allow distributed and multi‐user access to data and correlation of state with departmental events such as physician sign‐off. Conclusion: The immediate mode user‐interface solves these problems and results in strict management of data context and state and produces easily extendable code. The stated objectives implemented in CRADLE allow rapid deployment of advanced applications in radiation oncology.